The op amp connected in a negative feedback configuration, that is the o/p connected to the inverting terminal, tried to produce the same voltage at the inverting node as we applied to the non-inverting terminal, no matter whatever be the o/p. This happens because the opamp has a very high differential gain. so only way to not to clip or saturate the o/p in either positive or negative side is to make their both i/ps at the same potential. so when you apply the i/p to the inverting terminal and you connect the feedback to the inverting node, and your non-inverting terminal is grounded, then op amp tries to force the inverting node at the ground potential and whatever the difference between these two nodes are amplified.